five

STN PLAD (STN Power Line Assets Dataset)

收藏
OpenDataLab2026-05-24 更新2024-05-09 收录
下载链接:
https://opendatalab.org.cn/OpenDataLab/STN_PLAD
下载链接
链接失效反馈
官方服务:
资源简介:
STN PLAD 是多个高压电力线组件的高分辨率和真实世界图像数据集。它有 2,409 个带注释的对象,分为五类:输电塔、绝缘体、垫片、塔板和 Stockbridge 阻尼器,它们的大小(分辨率)、方向、照明、角度和背景各不相同。属性 图像大小:5472×3078 或 5472×3648 总图像:133 总实例:2409 每张图像的平均实例数:18.1 对象类别数(不同资产):5 其他统计数据:摘要 许多电力线公司正在使用无人机进行检查例如,通过让他们的工人爬上高压电力线塔,而不是让他们的工人处于危险之中。检查的一项关键任务是检测和分类输电线路中的资产。然而,与电力线资产相关的公共数据稀缺,阻碍了该领域的更快发展。这项工作提出了电力线资产数据集,其中包含多个高压电力线组件的高分辨率和真实世界图像。它有 2,409 个带注释的对象,分为五类:输电塔、绝缘体、垫片、塔板和 Stockbridge 阻尼器,它们的大小(分辨率)、方向、照明、角度和背景各不相同。这项工作还对流行的深度目标检测方法进行了评估,显示出相当大的改进空间。基线结果 mAP:89.2% 资产平均精度 输电塔 0.900 绝缘子 0.894 隔板 0.856 塔板 0.971 Stockbridge 阻尼器 0.838 平均 0.892

STN PLAD is a high-resolution real-world image dataset for multiple high-voltage power line components. It contains 2,409 annotated objects categorized into five classes: transmission towers, insulators, spacers, tower plates, and Stockbridge dampers, with variations in size (resolution), orientation, lighting, shooting angle and background. Attributes: Image size: 5472×3078 or 5472×3648 Total images: 133 Total instances: 2409 Average number of instances per image: 18.1 Number of object categories (distinct assets): 5 Summary: Many power line companies use drones for inspection, replacing the dangerous practice of having workers climb high-voltage power line towers. A key task in power line inspection is detecting and classifying assets in transmission lines. However, public data related to power line assets is scarce, hindering faster progress in this field. This work presents the power line asset dataset, which includes high-resolution real-world images of multiple high-voltage power line components, with 2,409 annotated objects divided into five classes: transmission towers, insulators, spacers, tower plates, and Stockbridge dampers, with varying sizes (resolutions), orientations, lighting conditions, angles and backgrounds. This work also evaluates popular deep object detection methods, showing considerable room for improvement. The baseline results are as follows: mAP: 89.2%; Average precision per asset: transmission towers 0.900, insulators 0.894, spacers 0.856, tower plates 0.971, Stockbridge dampers 0.838; overall average 0.892.
提供机构:
OpenDataLab
创建时间:
2022-05-30
搜集汇总
数据集介绍
main_image_url
以上内容由遇见数据集搜集并总结生成
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作